STA 364 Final Exam

Spring 2025

Setup and Data

library(fpp3)

Registered S3 method overwritten by 'tsibble':
  method               from 
  as_tibble.grouped_df dplyr

── Attaching packages ──────────────────────────────────────────── fpp3 1.0.1 ──

✔ tibble      3.2.1     ✔ tsibble     1.1.5
✔ dplyr       1.1.4     ✔ tsibbledata 0.4.1
✔ tidyr       1.3.1     ✔ feasts      0.4.1
✔ lubridate   1.9.4     ✔ fable       0.4.1
✔ ggplot2     3.5.1     

── Conflicts ───────────────────────────────────────────────── fpp3_conflicts ──
✖ lubridate::date()    masks base::date()
✖ dplyr::filter()      masks stats::filter()
✖ tsibble::intersect() masks base::intersect()
✖ tsibble::interval()  masks lubridate::interval()
✖ dplyr::lag()         masks stats::lag()
✖ tsibble::setdiff()   masks base::setdiff()
✖ tsibble::union()     masks base::union()

old_ny_childcare <- ny_childcare |> filter(year(Month)<2020)

Questions

1. Using appropriate plots, discuss which ETS models might be best and why. Match the components (E,T,S) and their functional form (Additive, Multiplicative) with what you see in your plots.

2. Using appropriate plots, discuss which ARIMA, or SARIMA models might be best. References specific plots and be specific.

3. Fit all appropriate models, including those you just mentioned; we have covered during the course using a training and testing approach, holding the last 2 years for testing.

4. Compare all models using multiple accuracy metrics on the training and testing set and visualize their forecasts on the testing set (and the whole series). Pick the best 2.

5. Assess the quality of the two chosen models and select the best overall model.

6. Using the best model, refit that type of model on the complete data and provide forecasts (plot) for 2020. Do you think your forecast is accurate? Why or why not?